[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#1000977: lintian: bogus elf-error in debug symbols



Control: tags -1 + patch
Control: forwarded -1 https://salsa.debian.org/lintian/lintian/-/merge_requests/387

On Thu, 02 Dec 2021 at 02:58:56 +0100, Andreas Beckmann wrote:
> W: hello-dbgsym: elf-error In program headers: Unable to find program interpreter name [usr/lib/debug/.build-id/69/1f9c7ce3fe763f9a6ddc8269eb2828af1ec0ee.debug]

I think Felix is correct to say that the root cause #1000449 is really
a binutils bug: if it's legitimate to have an ELF binary contaning only
detached debug symbols, then readelf shouldn't complain about that.

However, for binutils it's a minor or wishlist bug, whereas for
Lintian I'm finding that it significantly harms Lintian's usefulness
as a QA tool (at least for my uses of Lintian): when every package with
compiled binaries ends up with multiple W-level tags, it makes Lintian's
signal-to-noise ratio a lot worse, which makes it more likely that
uploaders will not notice a regression that Lintian could have warned
them about.

I think it would be proportionate for Lintian to work around #1000449
by ignoring this particular readelf error, for example with the change
I proposed in
<https://salsa.debian.org/lintian/lintian/-/merge_requests/387>.
Since #1000449 has already been reassigned to binutils, perhaps we can
use #1000977 to represent the request for a workaround?

    smcv


Reply to: